[Beowulf] how fast can NFS run?

Joel Jaeggli joelja at darkwing.uoregon.edu
Tue Jan 31 22:35:34 PST 2006

On Tue, 31 Jan 2006, Bruce Allen wrote:

>I'd like to know the fastest that anyone has seen an NFS server run, over
>either a 10Gb/s ethernet link or a handful of link aggregated
>(channel-bonded) Gb/s ethernet lines.

The netapp r200 we have serving user data pushes a peak of around 150MB/s, 
it has periods where the ten minute average will approach 100MB/s.  Qith a 
similar sized r200 and a smaller number of clients in the POC lab at 
netapp while evaling the hardware we got the thing to deliver order of 
220MB/s. This is 56 x 300GB drives. certainly they sell bigger faster 
filers than the ones we presently have.

>This would be with a small number of clients making large file sequential
>reads from the same NFS host/server.  Please assume that the NFS server 
>has 'infinitely fast' disks.

Disks aren't infinitely fast and do play a substantial part in the whole 
problem space, frankly if replacing them with silicon was cost 
effective I would, as it is we use bigger slower disks for almost all 
of our file server applications, just because we need the capacity/cost 
equation to balance out...

There are a lot of impediments to decent nfs performance, if you have an 
issue you get to play hunt the bottleneck, test, tune and repeat.

>I am told by one vendor that "NFS can't run faster than 100MB/sec".  I 
>understand or believe this.
>  If the server's local disks can read/write 
>at 300MB/s and the networking can run substantially faster than 100 MB/s, 
>I don't see any constraint to faster operation.  But perhaps someone on 
>this list can provide real-world data (or say why it can't work).

Joel Jaeggli  	       Unix Consulting 	       joelja at darkwing.uoregon.edu
GPG Key Fingerprint:     5C6E 0104 BAF0 40B0 5BD3 C38B F000 35AB B67F 56B2

More information about the Beowulf mailing list